Spark, by +Jason Pitre of Genesis of Legend, is a collaborative game of beliefs.
In many ways, that isn't just the very short soundbite for the game, it is the game. In a good way. Everything in Spark is directed at creating beliefs, for the characters and the setting, then setting about questioning those beliefs. When a situation arises that challenges your belief, do you confirm that belief, or do you refute it? That is the essence of the game. While that is the core, it is the rest of Spark which supports it
There is a GM and player division, though it is not a standard one. The GM may be best considered the steward and representative of the setting as a character, representing the faces of the various factions in play - in essence, the player representing the setting. The players, in turn, are given considerably greater responsibility for establishing scenes and moving the fiction and drama forward. It appears to work well as a hybrid between a traditional setup and a fully cooperative game.
Three example settings are presented in Spark, Neonippon a Shogunate Science Fiction, Quiet Revolution a Montreal Police Drama, and The Elemental Kingdom a Fantasy Under Siege. All of these are interesting and provide different examples of what can be done; Neonippon serves as an on-going example throughout the text, but the game will be at its strongest when working with your own setting that your group has investment in and serves to explore their interests.
Creating a setting is a straightforward collaborative procedure, though I hesitate at calling it "simple". This has nothing to do with the structure that is presented, and everything to do with getting everyone a the table to contribute and commit. There is also some good advice about making boundaries explicit. It is nice to see this front and center in a game that is about pushing beliefs and potentially
After you develop the basic premise of the setting, the group will create the setting beliefs that will drive the action and drama within the setting. Since they are playing the setting, the GM will select three of the proposed beliefs. Following this, each person (GM and the players) will create a faction for the setting, then create the mission statement for that faction. These must either confirm or refute one of the setting beliefs, and should give a general indication of where their goals and motivations lie.
Next, each faction will get a face. This is an NPC that represents that faction within the game (though other NPCs from those factions may show up). The faces will get the most development of any NPC and only the GM can portray then in a scene. After that, the relationships between the various factions are established and finally the agendas for each faction.
Agendas are specific activities that the various factions want to accomplish. They need to be within the grasp of that faction and something they could reasonably do within the time frame of the session. Because at the beginning of a session, each player will have a choice to make and one of the options is to prevent an agenda from succeeding; thus setting the stage for the events of the session. The other options are to choose a faction's next agenda, or create a new tie between two factions, or alter an existing tie.
Creating a PC is a simple prospect. There are attributes and talents, but most important are your beliefs. As with the setting, these are going to drive you personal actions and drama. In general, they should support or clash with a setting belief, or the same with another player's belief. This will serve to keep everything relevant and ripe for tension.
When actually playing the game, each of the scenes within a session is framed in a particular fashion. There is the platform, where the scene takes place, the tilt, what has triggered the scene, and the question, what is this scene about. One person will choose each of those, and anyone that doesn't get to make a selection has the option to portray major NPCs involved in the scene (just not any faces that may be present, those are only for the GM).
During the scene, the action will proceed through declarations. These can be statements about the scene, facts, or an action that a character is taking. If any other participant doesn't agree with your declaration and wants one different one, then you have a conflict. Whoever wins a conflict gets to use their version of reality for the events that actually take place. Any participant can spend their resources to help them win a conflict, but the winner of a conflict will always pay a cost.
After the question for a scene has been answered, the scene closes. From there, everyone will evaluate what beliefs were examined during the scene. The conflicts and collaboration that took place during the scene will determine this. By challenging beliefs, you earn influence and when you earn influence from all three of your beliefs, then everyone participating will earn an influence.
Influence can be spent to help win a conflict, or as the cost for victory. It can also be spent inspire a PC to change one of their beliefs. When that happens, both characters involved get an additional talent. When you get to the point that every PC has changed a belief, one of the setting beliefs will change, which will bring an attribute boost to everyone.
All of this is ultimately going to encourage everyone playing to be invested in all of the other characters (including the setting) at the table. Advancement is intrinsically related to everyone challenging their beliefs and working together as the people playing the game to move the story forward. It is a very collaborative style of play where the best results are going to come from setting up others and shining a spotlight on their beliefs and ensuring that everyone contributes.
This style of procedural play isn't going to appeal to every gamer out there. Every scene has a great deal of structure surrounding it, which means that every scene will have a discrete purpose and end, but there may not be a strong narrative link between the scenes. As well, the level of drama and tension built up within a scene can be easily diffused by the meta-game activities between scenes. This can be both a good thing and a bad thing.
Good in that it forces everyone to take a step back and examine the events that just took place from an objective point of view; seeing them without any personal investment. Given the charged nature the game intends to evoke, this could be important. It is problematic because all of the momentum that was built in the previous scene may be lost by breaking that tension, or by having the subsequent scene be unrelated to the previous events. It is a situation that will vary from group to group to a large degree.
In the end, this game is at its strongest when a group wants to explore some difficult and complicated concepts. Spark provides an excellent framework to examine different perspectives and build some meaningful discussion through this medium. This isn't really a pick-up or a beer and pretzels game. To get the most of of Spark, everyone will be best served by investing in a setting they create together, then populating it with characters they find interesting and asking questions they find compelling. This is going to be a rather personal game, for better or worse.